I simply do not trust generative AI answers from any platform. Even when there are sources shown, I can't know what parts of the answers are from which of the sources, or how the data from those sources might have been massaged, mangled, and misinterpreted to generate the answer.

I might as well just go directly to the sources, which is exactly what I would have done via blue links before (for example) #Google AI Overviews existed.

When I see (or hear, e.g., via Google Home) a conventional sourced answer to a question (e.g., According to the American Medical Association website ...) I know what I'm dealing with.

Not so with generative AI answers. They are impossible to reasonably evaluate. They are blacker boxes than the AI firms themselves, and the firms will not stand behind the answers and take any responsibility for them at all, hiding behind "these may be wrong, double check!" disclaimers.

They are, essentially, worse than useless.

L